Re: EARL Identifing The Error

This is a good approach, but in many cases you can simplify it by generating
more types of xpath.

for example, "html//img[@src='filename']" is one possibility. Matching an img
src attribute to a particular longdesc attribute is another self-contained
trick.

When an element has an id attribute (or you have access to the page and can
give it one) you can use that pretty easily.

Jim Ley and Nick Kew did some work on fuzzy pointers, based on different ways
of describing the content by taking a hash of certain features as an
identifier - so that you can change the content, but not the element types,
or vice versa...

I think extending the annotea context approach, so it doesn't have to have an
xpointer value but could refer to a piece of RDF, is the best approach - in
its simple form using the xpointer is straightforward, and there is no
restriction on adding new types of pointer, except maintaining
interoperability. (This is already an issue - different tools have different
approaches today)

Cheers

Chaals

On Thu, 15 Jan 2004, Chris Ridpath wrote:

>
>When an accessibility error is detected in a document you need to identify
>the thing that caused the error. For example the particular image that has
>no ALT text.
>
>(Note: The thing that caused the accessibility problem I'll refer to below
>as the 'item'.)
>
>We've discussed this a bit and Charles has suggested that we use an Annotea
>context property[1].
>
>Using an Xpath expression is the most accurate method of describing an item
>but it has poor persistence. Make a small change at the beginning of a
>document and all the following Xpath expressions may be invalid.
>
>There are 2 (at least) cases when you need to identify an item:
>
>1) You inform the user that there is a problem with an item.
>2) User has made a decision on an item (e.g. image does not require a
>longdesc).
>
>When you tell the user that there is a problem with an item you can use an
>Xpath expression because you are referring to the document right now.
>
>When the user makes a decision on an item, you want that decision to remain
>even after the document has changed. For example, if the user moves the
>image up or down in the page its longdesc requirement stays the same. Using
>an Xpath expression won't work for identifying the items of past decisions.
>
>Here's a suggestion for a practical, but fuzzy, solution to identifying
>items in the document. It offers good persistence but may have problems with
>accuracy:
>
>Every item may be identified by its element name and an identifier string.
>The identifier string is made up from a unique characteristic of the item.
>
>For images the identifier string would be the SRC attribute.
>
>Example: element "img" and identifier "rex.jpg". You can then find the item
>within the document using this information.
>
>Note that there may be more that one item in the document that fits this
>pattern. In this case they would all be treated the same. The decision (does
>not require a longdesc) pertaining to the original item would apply to all
>of these.
>
>I've started using this system and it's been working out OK so far. I'm
>placing this info in my EARL statements as 'element', 'item' and 'xpath'
>attributes of the earl:Subject like this:
>
>    <earl:Assertion>
>        <earl:Subject element="IMG" identifier="images/header.gif"
>            rdf:resource="#subject"
>xpath="/HTML/BODY/TABLE/TR/TD/TABLE/TR[1]/TH/IMG"/>
>        <earl:Testcase
>rdf:resource="http://tile-cridpath.atrc.utoronto.ca/acheck/checks/checks.htm
>l#check8">
>            <earl:testId rdf:resource="8"/>
>            <earl:message>image does NOT require a long
>description</earl:message>
>        </earl:Testcase>
>        <earl:result
>rdf:resource="http://www.w3.org/WAI/ER/EARL/nmg-strawman#Pass"/>
>    </earl:Assertion>
>
>I'm interested in hearing comments or suggestions.
>
>[1] http://lists.w3.org/Archives/Public/w3c-wai-er-ig/2003Dec/0002.html
>
>Cheers,
>Chris
>

Charles McCathieNevile  http://www.w3.org/People/Charles  tel: +61 409 134 136
SWAD-E http://www.w3.org/2001/sw/Europe         fax(france): +33 4 92 38 78 22
 Post:   21 Mitchell street, FOOTSCRAY Vic 3011, Australia    or
 W3C, 2004 Route des Lucioles, 06902 Sophia Antipolis Cedex, France

Received on Monday, 26 January 2004 07:32:40 UTC